This publication was produced on behalf of the Australian Stroke Clinical Registry (AuSCR) Consortium partners and

was approved by the AuSCR Steering Committee.

Suggested citation:

Breen S, Cadilhac DA, Lannin NA, Kim J, Dalli L, Anderson CS, Kilkenny M, Faux S, Dewey H, Hill K, Donnan G, Grimley R,

Campbell B, Mitchell P, Middleton S on behalf of the AuSCR Consortium. The Australian Stroke Clinical Registry Annual

Report 2019 Volume 2: Patient Outcomes. The Florey Institute of Neuroscience and Mental Health; July 2021, Report

No. 12, 22 pages.

Any enquires about, or comments on, this publication should be directed to:

Administration

Australian Stroke Clinical Registry Office

The Florey Institute of Neuroscience and Mental Health

245 Burgundy Street

Heidelberg, Victoria 3084, Australia

Ph: +61 3 9035 7264

Fax: +61 3 9035 7304

Email: admin@auscr.com.au

Website: www.auscr.com.au

@AustStrokeReg

July 2021

Report No. 12

Consortium partners:

CONTENTS EXECUTIVE SUMMARY ................................................................................................................................................ 4

INTRODUCTION .......................................................................................................................................................... 5

Overview of post-discharge outcomes data analysEs ................................................................................................. 6

POST-DISCHARGE HEALTH OUTCOME INFORMATION ................................................................................................ 8

Response rates ............................................................................................................................................................ 8

Follow-up data ............................................................................................................................................................ 9

Disability – modified Rankin Scale ............................................................................................................................ 10

Health-related quality of life ..................................................................................................................................... 13

Survival ..................................................................................................................................................................... 16

Risk-adjusted mortality ............................................................................................................................................. 17

Participation in research ........................................................................................................................................... 18

Unmet information needs ......................................................................................................................................... 18

DISCUSSION ...............................................................................................................................................................19

Follow-up data collection .......................................................................................................................................... 19

Patient outcomes ...................................................................................................................................................... 19

Conclusion ................................................................................................................................................................. 20

REFERENCES ..............................................................................................................................................................21

APPENDIX A: ABBREVIATIONS ...................................................................................................................................22

LIST OF TABLES

Table 1: Characteristics of patients with and without post discharge information ........................................................... 9

Table 2: Recurrent strokes, readmissions and living arrangements ................................................................................. 10

Table 3: modified Rankin Scale scores .............................................................................................................................. 11

Table 4: Quality of life assessment among survey respondents, stratified by stroke type .............................................. 14

LIST OF FIGURES

Figure 1: Follow-up of acute events in 2019 (excluding paediatric events) ....................................................................... 8

Figure 2: Unadjusted modified Rankin Scale score by stroke type at follow-up .............................................................. 12

Figure 3: Unadjusted modified Rankin Scale score for patients receiving endovascular clot retrieval at follow-up ....... 12

Figure 4: Unadjusted proportion of patients reporting problems with the EQ-5D-3L .................................................... 13

Figure 5: Health-related quality of life scores according to the adjusted VAS at follow-up, by hospital ......................... 15

Figure 6: Cumulative hazard of death, by treatment in a stroke unit (ischaemic stroke) ................................................ 16

Figure 7: Cumulative hazard of death, by treatment in a stroke unit (intracerebral haemorrhage) ............................... 16

Figure 8: 30 day Risk-adjusted mortality for ischaemic stroke (excluding transfers) ....................................................... 17

Figure 9: 30-day Risk-adjusted mortality for intracerebral haemorrhage (excluding transfers)..……………………………….…17

Figure 10: 30 day Risk-adjusted mortality for ischaemic stroke (including transfers) ...................................................... 18

Figure 11: 30-day Risk-adjusted mortality for intracerebral haemorrhage (including transfers)..…….……………………….…18

EXECUTIVE SUMMARY

The Australian Stroke Clinical Registry (AuSCR) provides a standardised approach for

hospitals to monitor, review and improve stroke care in compliance with national

acute stroke care standards. This report outlines the patient outcomes for 2019

stroke and transient ischaemic attack (TIA) admissions and is intended to be read in

conjunction with the AuSCR 2019 Annual Report Volume 1: Acute Care Provision.

AuSCR participation in 2019 and the collection of patient-reported outcomes • In 2019, 72 hospitals contributed data to the AuSCR (40% from Victoria, 29% from Queensland, 19% from New

South Wales, 4% each from Tasmania and South Australia and 3% from the Australian Capital Territory).

• This report is intended to be read alongside the AuSCR 2019 Annual Report Volume 1: Acute Care Provision

(https://auscr.com.au/about/annual-reports/).

• At 90-180 days after admission, collection of patient-reported outcomes was completed for 10,413 people who

experienced a stroke or transient ischaemic attack (TIA) aged over 18 years. This represented 71% of all eligible

episodes for 2019.

• Responders: were predominantly male (57%), had a median age of 72 years and had a median time to follow-up

survey completion of 145 days following admission for stroke or TIA.

Patient-reported outcomes obtained 90-180 days following admission • Most patients were living at home (85%). Compared to patients with ischaemic stroke, patients with intracerebral

haemorrhage were less likely to report living at home (OR: 0.57; 95% CI: 0.48-0.69), while patients with TIA were

more likely to report living at home (OR: 1.67; 95% CI: 1.37-2.04).

• Approximately 5% of patients had experienced a recurrent stroke, and 23% reported hospital readmissions.

• Almost half of respondents had a score of 0 (25%; representing no disability or symptoms) or 1 (22%; representing

no significant disability despite symptoms) on the modified Rankin Scale. Fewer patients with intracerebral

haemorrhage (ICH) reported no symptoms at follow up (13%) in contrast to ischaemic stroke (22%), undetermined

stroke type (28%) or TIA (45%).

• Health-related quality of life as measured by the EuroQol Five Dimensions Three Levels (EQ-5D-3L) survey

indicated that: 48% of patients reported problems with mobility; 31% with self-care; 57% with completion of usual

activities; 47% had pain/discomfort; and 46% had anxiety or depression. Patients with ICH reported more

problems on all of the five dimensions of quality of life than patients with other types of stroke or TIA.

• Self-reported overall health: The mean Visual Analogue Scale (VAS) score, which represents worst (0) and best

(100) imaginable health, was 69 compared to the reported population norm of 83. Mean VAS scores were lowest

for ICH (mean 63) and greatest in patients with undetermined stroke (mean 73).

• The overall minimum and maximum adjusted mean VAS scores differed between hospitals by almost 23 points

highlighting variation in outcomes depending on which hospital a patient was treated in.

Survival outcomes • Of the total acute stroke events recorded in 2019, 9% resulted in death prior to hospital discharge, 6% within 90

days of admission, and 2% between 90 and 180 days of admission.

• Treatment in a stroke unit was associated with a lower hazard of death at 180 days after admission than treatment

in an alternate ward for patients with either ICH or ischaemic stroke.

Other

• Almost half (49%) of respondents requested more information about stroke at 90-180 days after admission.

• 62% of respondents agreed to be contacted for future research study opportunities.

INTRODUCTION The Australian Stroke Clinical Registry (AuSCR) is a collaborative national effort to

monitor and support improvements to the quality of acute care for patients who

have a stroke and transient ischaemic attack (TIA). Since 2009, the AuSCR has

facilitated the collection of national data on consecutive patients admitted to

hospital with acute stroke or TIA, which has been used to inform improvements to

the health system.1

The AuSCR adheres to the national guidelines for best-

practice in clinical quality registries,2 and can be used

by both public and private hospitals. Adult and

paediatric cases are included. All participating

hospitals are required to have ethics and hospital-

specific governance approvals. As recommended for

national registries, an ‘opt-out’ model for patient

inclusion is used,3 in addition to a waiver of consent

for people who die while in hospital.

In the AuSCR, data are collected on the provision of

evidence-based therapies, supplemented with clinical

and demographic patient information, to provide an

indication of the quality of acute stroke care received.

Data are collected in the Australian Stroke Data Tool

(AuSDaT) which is a harmonised online platform

enabling standardised and systematic collection for

multiple national stroke data collection programs.

AuSCR data collection programs facilitate the

collection of data for patients with stroke/TIA who are

admitted to participating hospitals, and also patients

presenting to Emergency Departments (ED) prior to

transfer for continued acute care at another hospital.

The combination of programs enables the AuSCR to

fully document the acute care treatment pathway for

patients with stroke or TIA. These programs also

enable the collection of the original national minimum

processes of care dataset for assessing quality of care.2

For more information on acute data collection in the

AuSCR, and acute care provision in 2019, please refer

to the 2019 Annual Report Volume 1: Acute Care

Provision (https://auscr.com.au/about/annual-

reports/).

Patient-reported outcomes data are obtained via a

standardised survey administered at 90-180 days after

admission (also referred to as patient follow-up). In

addition to objective outcomes, including subsequent

stroke, hospital readmissions and current living

arrangements, subjective outcomes including self-

reported functional disability and health-related

quality of life are collected, namely the modified

Rankin Scale (mRS)4 and the EuroQol Quality of Life

measure (EQ-5D-3L).5 Function and health-related

quality of life is also collected from paediatric patients.

Paediatric patients use the Pediatric Quality of Life

(PedsQL)6 Inventory Measurement Model which is a

modular approach to measuring health-related quality

of life in children and adolescents.

The AuSCR office staff are responsible for following up

registrants who have not: been reported as deceased;

previously refused follow-up; or opted out of the

registry. Patient-reported outcomes data are collected

via a paper survey or telephone. A postal survey with a

reply paid envelope is mailed to each patient to collect

post-discharge information. If the surveys are not

returned within six weeks, another set of surveys are

posted to the patient and/or next of kin. After another

six-week period, a comprehensive telephone attempt

to the patient (and next of kin) is conducted. Patient

data collected via mail or telephone are entered into

the AuSDaT.

If all forms of contact are unsuccessful, then the

patient is deemed lost to follow-up, if survivor status is

not verified. Survival status is determined via annual

data linkage with the National Death Index (NDI) made

available by the Australian Institute of Health and

Welfare.

In 2019, we received follow-up data for 10,413

admitted episodes of patient care representing 18,733

individual patients from 72 Australian hospitals. The

data presented in this report provide insights into the

health outcomes of the 2019 registrants in AuSCR.

This report is intended to be read in conjunction with

the AuSCR 2019 Annual Report Volume 1: Acute Care

Provision (https://auscr.com.au/about/annual-

reports/).

6

OVERVIEW OF POST-DISCHARGE OUTCOMES DATA ANALYSES

Unless otherwise stipulated, patient data collected at

90-180 days post admission (ie follow-up data), were

analysed using descriptive statistics and multivariable

logistic regression, with adjustment for patient case

mix using age, sex, stroke type, ability to walk on

admission (stroke severity), in-hospital stroke, and

patient transfer from another hospital, as appropriate.

All episodes were eligible for follow-up except:

• where acute care data were not entered by

the hospital within 180 days post-admission;

or

• patients who were reported as deceased

prior to the 90-180 day follow-up.

In the case where there was a second admission

within 180 days of the first admission, follow-up was

only completed for the first admission.

Where comparisons are presented with large sample sizes, the reader should consider that the clinical relevance related to proportional differences may be more important than the statistical significance.

Functional disability is measured using the modified

Rankin Scale (mRS) which is a standardised instrument

used to categorise deficits after stroke.4 It provides a

simple descriptive profile across six levels including: 0

(no symptoms at all); 1 (no significant disability

despite symptoms; 2 (slight disability); 3 (moderate

disability); 4 (moderately severe disability) and 5

(severe disability). The mRS analyses in this report do

not include values of 6 (to signify death) for patients

who were deceased at the time of follow-up.

Registrants were asked to rate their level of functional

disability on the day of completion of the follow-up

survey.

Multivariable ordinal analysis of mRS data was

performed using partial proportional odds models to

assess changes across the scores between 0 to 5. As

recommended by the European Stroke Organisation

(https://eso-stroke.org/outcome-measures-stroke-

modified-rankin-scale-ordinal-logistic-regression/),

these analyses were adjusted for patient case mix

differences, clustering by hospitals, and presented as

common odds ratios. Common odds ratios can be

interpreted as the odds of a patient shifting to a higher

(or more severe) category of the mRS. For example, a

common odds ratio of 0.80 for a treatment would

mean that for a patient with an mRS of 4, the

treatment would increase a patients chance of

reporting a mRS score of 0 to 3 by 20%.

Health-related quality of life (HRQoL) is measured in

the AuSCR using the European Quality of Life measure

of health status (EQ-5D™)5, specifically the three-level

version of the instrument EQ-5D-3L. The EQ-5D-3L is a

standardised survey instrument used as a generic

measure of health outcome (see

http://www.euroqol.org/). It provides a simple

descriptive profile across five dimensions: mobility,

self-care, usual activities, pain and discomfort, and

anxiety and depression. Each dimension is divided into

three levels: no problems (1), some or moderate

problems (2) and extreme problems (3). Responses to

the EQ-5D-3L were dichotomised and reported based

on whether registrants had ‘any’ (ie moderate or

extreme) versus ‘no’ problems with each dimension.

Additionally, the EQ-5D-3L includes a self-rated

summary score of overall health using a Visual

Analogue Scale (VAS). The VAS responses range from

zero (worst imaginable health state) to 100 (the best

imaginable health state). Responses to the VAS were

described using the mean (and standard deviation) to

facilitate comparisons with published estimates for

the normative population. For hospital comparisons,

mean VAS scores were standardised to adjust for

differences in case mix using methodology similar to

that described below for mortality. Where possible,

reference to the median (and interquartile range) was

also made to assist in the interpretation of results.

Other patient-reported outcomes such as subsequent

stroke(s), hospital readmissions, current living

arrangements, need for stroke information and

interest in future research participation are captured

with single item standalone questions.

Case mix adjusted survival is measured using data

obtained from linkage to the NDI. Analysis of deaths

up to 180 days following admission was performed for

those who had experienced an episode of care in

2019. Cox proportional hazards regression analysis

was also performed on survival status for those who

received stroke unit care, compared with those who

did not. The survival analysis was adjusted for age, sex,

stroke type, in-hospital stroke, transfer from another

hospital and ability to walk on admission.

7

Risk-adjusted mortality rate (RAMR) was calculated

for each hospital based on admitted patients who had

died within 30 days of their admission. Separate RAMR

were calculated for patients with ischaemic stroke and

intracerebral haemorrhage (ICH). To ensure reliable

estimates, analyses were conducted for individual

hospitals that provided at least 200 episodes for cases

of ischaemic stroke, and at least 50 episodes for ICH in

2019. For hospitals with fewer episodes, data from

2018 and 2019 were pooled and used to derive

mortality estimates if the minimum number of

episodes was achieved across both years (i.e. at least

200 episodes for cases of ischaemic stroke, and at

least 50 episodes for ICH).

RAMR analyses were completed using two different

methods. The first method excluded patients who

were transferred from another hospital, in-hospital

strokes and patients with TIA. This method is

consistent with previous AuSCR Annual Reports

whereby deaths are summarised for patients treated

at only one hospital and are fully attributed to that

hospital.

The second RAMR method included patients who were

managed by more than one hospital for their episode

of stroke care (ie included all patients treated at an

individual hospital irrespective of whether they

presented directly or were transferred). In this

analysis, multiple episodes related to the same acute

stroke event were included, and mortality was

attributed equally to all hospitals involved in the

provision of patient care for that stroke event. RAMR

were calculated by dividing the risk-adjusted hospital

specific mortality by the risk-adjusted average hospital

mortality, and then multiplying by the unadjusted

proportion of deaths in the whole sample. Hospitals

with risk-adjusted mortality outside the three

standard deviation threshold limits are considered to

have ‘special cause variation’. All analyses were

performed using STATA/SE 16.0 (College Station, USA,

2019).

All RAMR models were adjusted for age, sex, country

of birth, Indigenous status, socioeconomic position,

stroke type, ability to walk on admission and previous

history of stroke. Socioeconomic position was based

on the Index of Relative Socioeconomic Advantage and

Disadvantage, an area-based measure of social and

economic circumstances published by the Australian

Bureau of Statistics. This measure of socioeconomic

position was determined for each patient using the

postcode of residence recorded in the AuSCR for the

acute episode.

8

POST-DISCHARGE HEALTH OUTCOME INFORMATION

RESPONSE RATES

Of the 14,728 adult acute events eligible for follow-up

by the AuSCR office, patients or their proxy provided

follow-up data for 10,413 (71%) acute events (Figure

1). Median time to follow-up was 145 days from

stroke admission (Q1 to Q3: 108 to 187 days).

There were 39 acute events occurring in patients

under the age of 18. Of these, 27 were treated in a

children’s hospital and 12 were treated in adult

hospitals. Of the total paediatric stroke events, 14

were ineligible for follow-up as 3 events were fatal, 1

event was within 180 days of a previous event, and 10

were registered in the AuSCR after 180 days. Of the

remaining 25 paediatric events eligible for follow-up,

21 follow-ups were completed by patients or their

proxies. Three patients were aged less than 1 year,

eight were aged between one and 11 years, and ten

were aged between 12 and 18 years.

* Number of acute events differ slightly from values reported in Volume I of the 2019 AuSCR Annual Report as a separate data

extract was provided for this report in which additional data cleaning operations had been performed by the AuSCR office.

† Deaths after discharge and within 90 days of admission were determined using the NDI.

‡ Contact unable to be made using our follow-up protocol (two postal attempts and one telephone attempt)

§ Deaths between 91 and 180 days of admission were determined using the NDI. Note that this figure does not represent total

deaths between 91-180 days following admission as some deaths occurred after completion of the follow-up survey.

FIGURE 1: FOLLOW-UP OF ACUTE EVENTS IN 2019 (EXCLUDING PAEDIATRIC EVENTS)

FIGURE 2: FOLLOW-UP OF ACUTE EVENTS IN 2019 (EXCLUDING PAEDIATRIC EVENTS)

9

FOLLOW-UP DATA

Of the patients eligible for follow-up, differences were

identified in both demographic and clinical

characteristics, as well as in processes of care which

are the basis of our quality indicators (Table 1).

At follow-up, approximately 5% of patients had

experienced a recurrent stroke and almost one in four

registrants (23%) reported hospital readmissions

(Table 2). The most common reasons for readmissions

were for recurrent stroke or TIA (18%) or for

cardiovascular disease (16%; Table 2). There were no

significant associations between stroke type and

either recurrent stroke (P=0.56) or subsequent

hospital readmission (P=0.25).

Most registrants were living at home (85%), while 24%

of registrants were living alone. There were 1,124

registrants (11%) living in low- or high-level residential

care for whom information at follow-up was obtained.

Differences in living situation were observed based on

stroke type after adjustment for age, sex,

socioeconomic position, and ability to walk on

admission. Compared to patients with ischaemic

stroke, patients with intracerebral haemorrhage were

less likely to report living at home (OR: 0.57; 95% CI:

0.48, 0.69) while patients with transient ischaemic

attack were more likely to report living at home at the

time of follow-up (OR: 1.67; 95% CI: 1.37, 2.04).

TABLE 1: CHARACTERISTICS OF PATIENTS WITH AND WITHOUT POST DISCHARGE INFORMATION

Completed (N= 10,413)

Not completed (N= 4,315)

P Value

Age (years), mean (SD) 72 (13) 70 (15) <0.001

Female, n (%) 4,389 (43) 1,887 (45) 0.003

Aboriginal and/or Torres Strait Islander, n (%)

164 (2) 136 (4) <0.001

Type of stroke, n (%) Ischaemic Intracerebral haemorrhage Transient ischaemic attack Undetermined

7,555 (73)

848 (8) 1,776 (17)

163 (2)

2,954 (70) 446 (11) 764 (18)

68 (2)

<0.001

Able to walk on admission, n (%) 5,102 (52) 1,826 (46) <0.001

Length of hospital admission (days), median (Q1 to Q3)

4 (2 to 7) 4 (2 to 8) <0.001

Treated in a stroke unit, n (%) 8,534 (82) 3,351 (78) <0.001

Excludes paediatric episodes SD: standard deviation Q1: 25th percentile Q3: 75th percentile

10

TABLE 2: RECURRENT STROKES, READMISSIONS AND LIVING ARRANGEMENTS

N= 10,413

n (%)

Had a recurrent stroke 525 (5%)

Readmitted to hospital 2,331 (23%)

Time to readmission (days), median (Q1, Q3) 81 (42, 126)

Reason for readmission N= 2,321

Stroke or transient ischaemic attack 428 (18%)

Cardiovascular disease 364 (16%)

Elective surgery 243 (10%)

Injury 228 (10%)

Other neurological condition 163 (7%)

Infection 151 (7%)

Respiratory disease 132 (6%)

Gastrointestinal disease 112 (5%)

Other 500 (22%)

Location at time of follow-up interview N= 10,404

Home 8,840 (85%)

Living alone

Living with others

2,150 (24%)

6,636 (76%)

With care support

Without care support

4,262 (48%)

4,578 (52%)

Institutional care or other setting 1,497 (15%)

In hospital

Transitional care services

Low level care (hostel care)

High level care (nursing home)

Inpatient rehabilitation

Other

108 (7%)

97 (6%)

97 (6%)

1,027 (69%)

103 (7%)

65 (4%)

Missing data not included in denominators Excludes paediatric episodes

DISABILITY – MODIFIED RANKIN SCALE

The mRS is widely used in stroke studies to describe person-centred global disability outcomes in terms of the degree

of disability, or dependence, in daily activities with reference to pre-stroke activities. At follow-up, 25 percent of

registrants reported they were free from disability (mRS score = 0) and 22 percent had no significant disability despite

symptoms (mRS score = 1; Table 3).

The unadjusted median mRS score reported by patients with ICH was 3 (ie moderate disability), compared to a median

mRS score of 2 (ie slight disability, but can manage own affairs without assistance) for patients with ischaemic stroke,

and 1 (i.e. no significant disability despite symptoms) for patients with undetermined stroke or TIA (Figure 2).

11

In the adjusted ordinal analysis of the mRS (https://eso-stroke.org/outcome-measures-stroke-modified-rankin-scale-

ordinal-logistic-regression/) , patients with ICH were 85% more likely to report disability across increasing categories of

the mRS than patients with ischaemic stroke (common OR: 1.85; 95% CI: 1.62,2.12). In contrast, patients with

undetermined stroke type were 27% less likely to report disability across increasing categories of the mRS compared to

patients with ischaemic stroke (common OR: 0.73; 95% CI: 0.55,0.97). These estimates were adjusted for patient

clustering by hospital as well as age, sex, socioeconomic position, Aboriginal or Torres Strait Islander background, ability

to walk on admission, history of previous stroke, in-hospital stroke, hospital transfer status, management in a stroke

unit, and country of birth.

Of the 1,220 episodes where endovascular clot retrieval (ECR) was provided in 2019, mRS data were available for 729

patients with completed follow-up surveys. The mRS scores of these patients who received ECR are shown in Figure 3.

Of note is that people who are eligible for ECR have a more severe type of stroke associated with large vessel

occlusion in an artery of the brain, and this is why the proportion of patients reporting independence after treatment

will vary in contrast to all people who have experienced an ischaemic stroke.

TABLE 3: UNADJUSTED MODIFIED RANKIN SCALE SCORES AT FOLLOW-UP N=10,413

n (%)

Modified Rankin Scale

0 – No symptoms at all

1 – No significant disability despite symptoms

2 – Slight disability

3 – Moderate disability

4 – Moderately severe disability

5 – Severe disability

Missing

2,593 (25%)

2,297 (22%)

1,941 (19%)

2,155 (21%)

961 (9%)

374 (4%)

92 (1%)

12

mRS: modified Rankin Scale ICH: intracerebral haemorrhage TIA: transient ischaemic attack

mRS: modified Rankin Scale

FIGURE 2: UNADJUSTED MODIFIED RANKIN SCALE SCORE BY STROKE TYPE AT FOLLOW-UP

FIGURE 3: UNADJUSTED MODIFIED RANKIN SCALE SCORE FOR PATIENTS RECEIVING ENDOVASCULAR CLOT RETRIEVAL AT FOLLOW-UP

13

HEALTH-RELATED QUALITY OF LIFE

With respect to health-related quality of life, problems

were most frequently reported with completion of

usual activities (Figure 4). This was consistent across

all stroke types (Table 4). Patients with ICH reported

problems in all dimensions of the EQ-5D-3L more

commonly than the other diagnoses. Overall, 74% of

respondents reported problems with at least one

dimension of the EQ-5D-3L with this proportion being

greatest for those diagnosed with ischaemic stroke

(85%; Table 4).

The mean Visual Analogue Scale (VAS) score, which

represents a registrants’ self-reported overall health,

was 69 (median: 73; min-max: 0-100). Differences in

VAS scores were observed by stroke type, whereby

registrants with ICH had the lowest aggregate VAS

scores (mean: 63; median 67; Table 4) and registrants

with undetermined stroke had the greatest VAS scores

(mean: 73; median 79).

The VAS scores reported by AuSCR registrants should

be considered in relation to a mean of 83 reported for

the normative population of adults in the United

Kingdom.7 Compared to this reference population, the

VAS scores of AuSCR registrants were >10 points

worse after stratifying by age and sex (Table 4). As

these differences exceed 8 points, they represent a

clinically meaningful difference.8

The overall minimum and maximum adjusted mean

VAS scores differed between participating hospitals by

almost 23 points after adjustment for differences in

patient and clinical characteristics (Figure 5).

FIGURE 4: UNADJUSTED PROPORTION OF PATIENTS REPORTING PROBLEMS WITH THE EQ-5D-3L

14

TABLE 4: UNADJUSTED QUALITY OF LIFE ASSESSMENT AMONG SURVEY RESPONDENTS, STRATIFIED BY STROKE TYPE

Ischaemic N=7,555

ICH N=848

TIA N=1,776

Undetermined N=163

Proportion of patients reporting problems with each dimension of the EQ-5D-3L

Mobility 49% 61% 40% 41%

Self-care 31% 46% 21% 26%

Usual activities 58% 74% 43% 51%

Pain/Discomfort 47% 56% 42% 48%

Anxiety/Depression 46% 55% 41% 46%

Proportion of patients reporting problems with ≥1 dimension of the EQ-5D-3L

85% 75% 65% 75%

VAS mean (SD) 68 (22) 63 (23) 72 (20) 73 (19)

VAS median (Q1-Q3) 73 (50-85) 67 (50-80) 75 (60-89) 79 (60-85)

Missing responses (<2%) excluded from denominators ICH: intracerebral haemorrhage TIA: transient ischaemic attack Excludes paediatric episodes

15

Adjusted for differences in patient age, sex, socioeconomic position, country of birth, Indigenous status, type of stroke, in-hospital stroke, history

of previous stroke, and ability to walk on admission.

Mean of 83 for the normative adult population was obtained from the United Kingdom.7

The median number of responders to the VAS survey per hospital was 119 (min-max: 1-666).

FIGURE 6: HEALTH-RELATED QUALITY OF LIFE SCORES ACCORDING TO THE ADJUSTED VISUAL ANALOGUE SCALE AT FOLLOW-UP, BY HOSPITAL FIGURE 5: HEALTH-RELATED QUALITY OF LIFE SCORES ACCORDING TO THE ADJUSTED VAS AT FOLLOW-UP, BY HOSPITAL

16

SURVIVAL

Survival status was ascertained for the entire

AuSCR cohort using data linkage with the

National Death Index. Of the total acute

stroke events recorded in 2019, 9% of events

resulted in death prior to hospital discharge,

6% between discharge and 90 days of

admission, and 2% between 91 and 180 days

of admission. For patients with ischaemic

stroke, treatment in a stroke unit was

associated with a 50% lower hazard of death

at 180 days after admission than treatment

on an alternate ward (hazard ratio 0.50, 95%

confidence interval 0.45,0.55, p<0.001,

Figure 6). A larger effect of a 64% lower

hazard of death, was found for patients with

ICH (hazard ratio 0.36, 95% confidence

interval 0.31,0.41, p<0.001, Figure 7). These

analyses were adjusted for age, sex, ability to

walk on admission, in-hospital stroke and

transfer from another hospital.

FIGURE 6: CUMULATIVE HAZARD OF DEATH, BY TREATMENT IN A STROKE UNIT (ISCHAEMIC STROKE)

FIGURE 7: CUMULATIVE HAZARD OF DEATH, BY TREATMENT IN A STROKE UNIT (INTRACEREBRAL HAEMORRHAGE)

17

RISK-ADJUSTED MORTALITY

Risk-adjusted mortality rate comparisons at 30 days

following admission for ischaemic stroke and ICH

have been plotted for tertiary referral hospitals that

offer ECR services (yellow dots) and for other

hospitals that do not offer ECR services (black dots;

Figures 8 to 11).

To ensure reliable estimates, analyses were

conducted for individual hospitals that provided at

least 200 episodes of ischaemic stroke and at least

50 episodes of ICH in 2019. For hospitals with fewer

episodes, data from 2018 and 2019 were pooled

and used to derive mortality estimates if the

minimum number of episodes was achieved across

both years (ie at least 200 episodes of ischaemic

stroke and at least 50 episodes of ICH).

The risk-adjusted mortality rate for tertiary referral

hospitals have been colour-coded separately as

these hospitals are more likely to treat a different

cohort of patients with stroke, including more

complex and severe cases, than other hospitals

participating in the AuSCR.

For the first time in the evaluation of overall risk-

adjusted mortality, we have also analysed and

reported data for patients managed by more than

one hospital for their stroke event. In this analysis,

mortality was attributed equally to all hospitals

involved in the treatment of the acute stroke event

(Figures 10 and 11). The inclusion of hospital

transfers in RAMR calculations resulted in an

additional 3,182 episodes being included for

ischaemic stroke and 561 for ICH.

All RAMR models were adjusted for age, sex,

country of birth, Indigenous status, socioeconomic

position, stroke type, ability to walk on admission

and previous history of stroke. Combined 2018 and

2019 data were used to maximise the sample size

available for analysis, where required, at smaller

hospitals.

While there was some evidence of variation in RAMR

between hospitals, no hospitals exhibited special cause

variation (ie mortality more than three standard

deviations above the mean).

FIGURE 8: 30 DAY RISK-ADJUSTED MORTALITY FOR ISCHAEMIC STROKE, (EXCLUDING TRANSFERS)

FIGURE 9: 30 DAY RISK-ADJUSTED MORTALITY FOR INTRACEREBRAL HAEMORRHAGE (EXCLUDING TRANSFERS)

Each dot represents the risk-adjusted percentage of patients who were

deceased within 30 days of admission at each hospital.

Yellow dots represent large tertiary referral hospitals which provided ECR

services (ie hospitals participating in the AuSCR Black or Violet programs).

Black dots represent other hospitals which did not offer ECR services (ie

hospitals participating in the AuSCR Red, Navy or Green programs).

Excludes transfers, in-hospital strokes and hospitals with less than 200

episodes for ischaemic stroke or less than 50 episodes for intracerebral

haemorrhage.

18

For information on how to interpret funnel plots

please refer to page 7 of the AuSCR 2019 Annual

Report Volume 1: Acute Care Provision.

PARTICIPATION IN RESEARCH

Among the 10,117 registrants who answered the

question about whether they would be willing to be

contacted to participate in future research, 6,254

(62%) replied affirmatively. Compared to those who

did not reply in the affirmative, these patients were

younger (median age 73 vs 77 years, p<0.001) and

more often male (60% vs 54%, p<0.001).

UNMET INFORMATION NEEDS

Stroke can be a devastating and life changing event for

people and there is a possibility that stroke survivors

and their care providers have unmet care and

information needs. Given that the AuSCR protocol

includes a follow-up survey with survivors at 90-180

days post stroke or TIA, it presents an opportunity to

ask registrants whether they would like to receive

further information about stroke from the Stroke

Foundation. In 2019, 49% (n=4,990) of the 10,115

registrants who answered this question indicated that

they would like to receive an information package.

FIGURE 10: 30 DAY RISK-ADJUSTED MORTALITY FOR ISCHAEMIC STROKE (INCLUDING TRANSFERS)

FIGURE 11: 30 DAY RISK-ADJUSTED MORTALITY FOR INTRACEREBRAL HAEMORRHAGE (INCLUDING TRANSFERS)

Each dot represents the risk-adjusted percentage of

patients who were deceased within 30 days of

admission at each hospital.

Yellow dots represent large tertiary referral

hospitals which provided ECR services (ie hospitals

participating in the AuSCR Black or Violet

programs). Black dots represent other hospitals

which did not offer ECR services (ie hospitals

participating in the AuSCR Red, Navy or Green

programs).

Excludes in-hospital strokes and hospitals with less

than 200 episodes for ischaemic stroke or less than

50 episodes for intracerebral haemorrhage.

19

DISCUSSION This report outlines the patient outcomes for 2019 stroke and TIA admissions and is

intended to be read in conjunction with the AuSCR 2019 Annual Report Volume 1:

Acute Care Provision.

Follow-up data collection The overall rate of follow-up completion for eligible

registrants increased to 71% in 2019, from 69% in 2018

and 65% in 2017. This represents a very high rate of

completion when compared to other international

stroke registries such as SSNAP, which complete patient

follow-up on 30% of eligible patients.9 The trend of

increased rates of follow-up completion is a testament

to the hard work and refined protocols of the AuSCR

office. Ongoing training and support for hospital staff

are provided by the AuSCR office, with an increased

emphasis on timely data entry to enable follow-up

completion at 90-180 days post admission since 2017.

Patient outcomes This findings in this report illustrate high levels of

patients reporting difficulty across all five dimensions of

the EQ-5D-3L ranging from 31% for self-care to 57% for

difficulties in completing the usual activities of daily

living. The overall mean VAS score of 69 as a measure of

overall health (when compared to the normative

population mean value of 83) is also indicative of

ongoing health problems. Twenty-three percent of

patients documented readmissions at 90-180 days after

admission (with 18% of these for recurrent stroke or

TIA), and almost half (49%) of patients requested more

information about stroke. To address these findings,

there is a pressing need to consider improved models of

care and support for patients with stroke following

hospital discharge. National Registries, such as the

AuSCR, are ideally placed to play an important role in

the development and monitoring of such models of

care, where routinely collected follow-up data could be

used to identify patients with high levels of health

needs, or those at risk of poor outcomes (eg

readmissions). Summaries of individual patient data

could then be fed back to participating hospitals and

this information could prompt an outpatient review of

those most in need and or at risk of poor outcomes. In

the year ahead, the AuSCR will aim to build upon

previous experience with summarising patient-reported

outcome data7 to develop and test the effectiveness of

new models of care to improve patient outcomes after

stroke.

Significant variation in case mix adjusted analyses of the

EQ-5D-3L VAS was identified between hospitals in the

2019 dataset. The difference in mean VAS scores

between the best and worst performing hospital was 23

points. The clinically meaningful difference in VAS

scores for patients with stroke has been reported by

previous authors as eight points7 and further analysis to

explore the factors contributing to these variations is

required. The results reported using the AuSCR 2019

dataset also suggest that the VAS data may play an

important role in ongoing hospital quality improvement

initiatives such as the, StrokeLink and Tasmanian

Community of Practice programs.

The proportion of registrants who received

endovascular clot retrieval and reported an unadjusted

mRS scores of 0 to 2 was 59%. This is an encouraging

finding when compared to the Highly Effective

Reperfusion Using Multiple Endovascular Devices

(HERMES) collaboration finding 46% (291/633) of

patients with a mRS score between 0 to 2 using pooled

data reported from multiple ECR interventional

studies.12 In the adjusted ordinal analysis of mRS data,

patients with ICH were more likely to report disability

across increasing categories of the mRS than the other

stroke types or TIA. These observations align with the

unadjusted mRS scores by stroke type reported using

AuSCR data previously

(https://auscr.com.au/about/annual-reports/).

For the first time in the AuSCR Annual Report series, the

risk-adjusted mortality analyses were presented using

different methods to analyse the 2019 data. This

included delineation of tertiary referral hospitals and

non-tertiary referral hospitals; and by reporting

statistical models that both excluded and included

patients who had been treated for their acute event by

more than one hospital. Our analytical team and

supporting Management and Steering Committee

20

representatives believe the delineation of tertiary

referral hospitals in this report enabled clearer

comparisons of mortality rates for hospitals that are

likely to have a greater number of complex and severe

patients. If we did not enhance the reporting of RAMR

calculations, we may not have been able to account for

all differences in our statistical models of case mix

adjustment.

The inclusion of patients who were managed in more

than one hospital (ie irrespective of whether they

presented directly to a hospital or were transferred for

care) for the RAMR analyses enabled a more realistic

assessment of RAMR for hospitals with high rates of

patient transfers. Large tertiary hospitals therefore have

a complete overview of mortality for their patient

cohort when compared with models that exclude

patient transfers. In general, the exclusion or inclusion

of transferred patients receiving care at more than one

hospital made very little difference to RAMR.

While it is reassuring to report that no hospitals

exhibited special cause variation in RAMR, it must be

acknowledged that there may be other confounding

factors which contribute to mortality rates (eg patient

co-morbidities, size of the infarct or occlusion) and that

the RAMR completed by the AuSCR can only adjust for

routinely collected variables. The NIHSS is commonly

used for case mix adjustment in the field of stroke.10

However, while the completion rate of the NIHSS

variable has improved significantly since 2016 (with only

2% missing data in 2019), 42% of baseline NIHSS scores

are still documented as ‘unknown’ (see AuSCR 2019

Data Quality Report). We are therefore unable to adjust

RAMR reliably using the NIHSS, and continue to use the

ability to walk on admission as a proxy for stroke

severity.11 The AuSCR will continue to work with

hospitals to improve the rate of NIHSS documentation,

and to identify barriers to completion of this

measurement in clinical practice. Improvements in the

methods for case mix adjustment for RAMR are also

being explored, so that fairer comparisons between

hospitals with similar patient cohorts can be included in

future reports.

Patient outcomes presented in this report may also

have been influenced by patient discharge destination,

with the authors of previous studies reporting

differences in patient outcomes for those discharged to

inpatient rehabilitation, versus those who were

discharge home or to a residential aged care facility.13

Associations between discharge destination and patient

outcomes such as mortality, subsequent readmissions,

and Health-related Quality of Life (HRQoL) will be

further explored in future AuSCR Annual Reports.

Conclusion Patient outcomes for AuSCR registrations in 2019 were

similar to those observed using 2018 data. The

continued levels of unmet health needs reported by

patients with stroke, and the demand for more

information about stroke following hospital admission,

support the development and testing of new post-

discharge models of patient care. The collection of

standardised patient-reported outcomes data at 90-180

days in the AuSCR, means this leading national registry

is in an ideal position to facilitate the identification of

patients with high levels of health needs, or those at risk

of poor outcomes, for further support by treating

hospitals. Future work will be undertaken to develop

and test novel models of post-discharge support for

patients with stroke.

21

REFERENCES

1. Cadilhac DA, Lannin NA, Anderson CS, et al. Protocol and pilot data for establishing the Australian Stroke Clinical

Registry. International Journal of Stroke. 2010;5:217-226.

2. Australian Commission on Safety and Quality in Health Care. Operating Principles and Technical Standards for

Australian Clinical Quality Registries. Sydney: ACSQHC, 2008 November

3. Evans SM, Loff B, Cameron PA. Clinical registries: the urgent need to address ethical hurdles. MJA. 2013;198:134-

135.

4. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJA, van Gijn J. Interobserver agreement for the assessment of

handicap in stroke patients. Stroke. 1988;19:604-607

5. The EuroQol Group. EuroQol-a new facility for the measurement of health-related quality of life. Health Policy.

1990;16:199-208

6. Varni JW, Seid M, Rode CA. The PedsQL: measurement model for the pediatric quality of life inventory. Med Care.

1999;37:126-139

7. Kind P, Dolan P, Gudex C, Williams A. Variations in population health status: results from a United Kingdom

national questionnaire survey. BMJ. 1998;316:736-741.

8. Kim SK, Kim SH, Jo MW, Lee SI. Estimation of minimally important differences in the EQ-5D and SF-6D indices and their utility in stroke. Health Qual Life Outcomes. 2015;13:32.

9. Sentinel Stroke National Audit Program, 2019, School of Population Health and Environmental Sciences, King’s

College London on behalf of the Intercollegiate Stroke Working Party, SSNAP Annual Public Report 2018. Available

athttps://www.hqip.org.uk/wp-content/uploads/2019/06/Ref-142-SSNAP-Annual-Report-FINAL.pdf

10. Fonarow GC, Pan w, Saver JL et al. Comparison of 30-day mortality models for profiling hospital performance in

acute ischaemic stroke with vs without adjustment for stroke severity. JAMA. 308 (3); 257-264

11. Counsell C, Dennis M, McDowall M, Warlow C. Predicting outcome after stroke and subacute stroke: development

and validation of new prognostic models. Stroke. 2002; 33 :1041-1047.

12. Goyal M, Menon BK, Van Zwam WH et al. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-

analysis of individual patient data from five randomised trials. Lancet. 2016; 3877:1723-1731.

13. Lynch EA, Labberton AS, Kim J et al. Out of sight, out of mind: long-term outcomes for people discharged home, to

inpatient rehabilitation and to residential aged care after stroke. Disability and Rehabilitation. 2020; DOI:

10.1080/09638288.2020.1852616

22

APPENDIX A: ABBREVIATIONS

AuSCR Australian Stroke Clinical Registry

AuSDaT Australian Stroke Data Tool

ECR Endovascular clot retrieval

ED Emergency Department

EQ-5D-3LTM European Quality of Life - five dimension three level instrument

ECR Endovascular clot retrieval

HRQoL Health-related quality of life

ICH Intracerebral haemorrhage

mRS Modified Rankin Scale

NDI National Death Index

NIHSS National Institutes of Health Stroke Scale

Q1/Q3 25th percentile/75th percentile

PROMs Patient-reported outcome measures

RAMR Risk-adjusted mortality rate

SD Standard deviation

SSNAP Sentinel Stroke National Audit Program

STELAR Shared Team Efforts Leading to Adherence Results

TIA Transient ischaemic attack

VAS Visual Analogue Scale